AOS中基于反馈机制的主动队列管理算法研究
发布时间：2018-08-13 12:05
【摘要】：为了适应航空航天技术的飞速发展,国际空间数据系统咨询委员会(CCSDS)推出了拥有灵活多样的数据处理业务的高级在轨系统(AOS)协议。随着空间数据量的越来越大,以及数据流的自相似性的影响,空间网络发生拥塞就变得不可避免。为了缓解网络的拥塞,本文提出了一种基于反馈机制的主动队列管理算法,实现对拥塞的有效控制。本文的主要研究的内容包括:首先,详细介绍了高级在轨系统以及网络流量的自相似性。网络业务流量拥有着自相似的特性,这意味着传统的泊松模型对自相似业务已经不适用。本文介绍了两种适应自相似性的业务模型:ON/OFF模型和FBM模型,并分别对它们建模和仿真。然后对处理网络拥塞现象有很好效果的RED算法进行详细介绍,借鉴该算法的思想对主动队列管理算法进行改进。使用ON/OFF模型模拟自相似流量,统计每个信源ON状态的持续时间,并求其平均持续时间。通过将求得的平均持续时间与设定的门限值进行比较,决定对数据包的丢弃情况,实现调控数据传输速率,缓解网络拥塞的目的。最后,考虑到在引入队列管理模块的AOS多路复用模型中,各个模块之间可以相互结合,实现系统在控制网络拥塞方面的进一步优化。在改进后的队列管理算法的基础上加入反馈机制,将虚拟信道调度模块的帧缓存区的紧张程度反馈给队列管理模块,从而得到性能更加优越的基于反馈机制的主动队列管理算法。该主动队列管理算法同时考虑到数据包缓存区的紧张程度和数据帧缓存区的紧张程度,并使用MATLAB软件搭建仿真平台进行仿真。通过评估仿真结果的各项性能指标,得出加入反馈调节的主动队列管理算法能够极大提高系统性能的结论。
[Abstract]:In order to adapt to the rapid development of aerospace technology, the International Space data system Advisory Committee (CCSDS) has launched the advanced on-orbit system (AOS) protocol with flexible and diverse data processing services. With the increasing amount of spatial data and the influence of self-similarity of data flow, congestion in spatial networks becomes inevitable. In order to alleviate network congestion, an active queue management algorithm based on feedback mechanism is proposed to effectively control congestion. The main contents of this paper are as follows: firstly, the paper introduces the self-similarity of high-level on-orbit system and network traffic in detail. Network traffic has the characteristic of self-similarity, which means that the traditional Poisson model is no longer suitable for self-similar traffic. In this paper, we introduce two kinds of business models that adapt to self-similarity, namely: one / off model and FBM model, and model and simulate them respectively. Then the RED algorithm which can deal with network congestion is introduced in detail and the active queue management algorithm is improved by using the idea of the algorithm. Using ON/OFF model to simulate self-similar traffic, the duration of each source on state is counted and its average duration is obtained. By comparing the average duration with the set threshold, the data packet dropping is determined, the data transmission rate is adjusted and the network congestion is alleviated. Finally, considering that in the AOS multiplexing model with queue management module, each module can be combined with each other to achieve further optimization in network congestion control. On the basis of the improved queue management algorithm, the feedback mechanism is added, and the tension of the frame buffer of the virtual channel scheduling module is fed back to the queue management module. Thus, an active queue management algorithm based on feedback mechanism is obtained. The active queue management algorithm considers the tension of data packet buffer and data frame buffer at the same time, and uses MATLAB software to build a simulation platform for simulation. By evaluating the performance indexes of the simulation results, it is concluded that the active queue management algorithm with feedback regulation can greatly improve the performance of the system.
【学位授予单位】：沈阳理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：V443.1;TP393.06
[Abstract]:In order to adapt to the rapid development of aerospace technology, the International Space data system Advisory Committee (CCSDS) has launched the advanced on-orbit system (AOS) protocol with flexible and diverse data processing services. With the increasing amount of spatial data and the influence of self-similarity of data flow, congestion in spatial networks becomes inevitable. In order to alleviate network congestion, an active queue management algorithm based on feedback mechanism is proposed to effectively control congestion. The main contents of this paper are as follows: firstly, the paper introduces the self-similarity of high-level on-orbit system and network traffic in detail. Network traffic has the characteristic of self-similarity, which means that the traditional Poisson model is no longer suitable for self-similar traffic. In this paper, we introduce two kinds of business models that adapt to self-similarity, namely: one / off model and FBM model, and model and simulate them respectively. Then the RED algorithm which can deal with network congestion is introduced in detail and the active queue management algorithm is improved by using the idea of the algorithm. Using ON/OFF model to simulate self-similar traffic, the duration of each source on state is counted and its average duration is obtained. By comparing the average duration with the set threshold, the data packet dropping is determined, the data transmission rate is adjusted and the network congestion is alleviated. Finally, considering that in the AOS multiplexing model with queue management module, each module can be combined with each other to achieve further optimization in network congestion control. On the basis of the improved queue management algorithm, the feedback mechanism is added, and the tension of the frame buffer of the virtual channel scheduling module is fed back to the queue management module. Thus, an active queue management algorithm based on feedback mechanism is obtained. The active queue management algorithm considers the tension of data packet buffer and data frame buffer at the same time, and uses MATLAB software to build a simulation platform for simulation. By evaluating the performance indexes of the simulation results, it is concluded that the active queue management algorithm with feedback regulation can greatly improve the performance of the system.
【学位授予单位】：沈阳理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：V443.1;TP393.06
【参考文献】
相关期刊论文 前10条
1 刘立士;李清凡;田野;张艳琴;;基于帧紧迫度的边界可移动虚拟信道调度算法[J];科学技术与工程;2014年17期
2 安红章;吴开均;任飞;;AOS链路多级安全保护技术探讨[J];通信技术;2013年12期
3 徐凌;刘嘉q ;李亮;;自相似网络流量Hurst指数估计算法[J];科学技术与工程;2013年20期
4 刘海泉;罗钧e，
本文编号：2180937
本文链接：https://www.wllwen.com/guanlilunwen/ydhl/2180937.html
